| Finished | ||||||
|---|---|---|---|---|---|---|
| Warreg | Pea | IBM-POWER-compatibility | diff | 4.0+0.04 | LLR: 2.99 (-2.94, 2.94) [-5.00, 0.00] Games: 18034 W: 4329 L: 4283 D: 9422 Ptnml(0-2): 255, 2133, 4183, 2203, 243 | different modules for avx2 and base, should fix issue. |
| Warreg | Pea | 4kb-threats-2epoch-session4 | diff | N=100000 | LLR: 2.95 (-2.94, 2.94) [0.00, 5.00] Games: 2408 W: 657 L: 527 D: 1224 Ptnml(0-2): 20, 248, 559, 336, 41 | session 3 passed, new net uploaded |
| Warreg | Pea | 4kb-threats-2epoch-session3 | diff | N=100000 | LLR: 3.03 (-2.94, 2.94) [0.00, 5.00] Games: 1482 W: 480 L: 334 D: 668 Ptnml(0-2): 21, 141, 289, 251, 39 | threat inputs session 3 vs session 2, 60m positions |
| Warreg | Pea | 4kb-threats-2epoch-session2 | diff | N=100000 | LLR: 2.99 (-2.94, 2.94) [0.00, 5.00] Games: 1092 W: 388 L: 243 D: 461 Ptnml(0-2): 14, 86, 228, 177, 41 | threat inputs session 2 vs session 1, 80m positions |
| Warreg | Pea | 4kb-threats-2epoch-session1 | diff | N=100000 | LLR: 3.05 (-2.94, 2.94) [0.00, 5.00] Games: 3440 W: 1204 L: 1033 D: 1203 Ptnml(0-2): 104, 353, 674, 446, 143 | threat inputs session 1 vs session 0 |
| Warreg | Pea | 4kb-threats-2epoch | diff | N=100000 | LLR: 2.95 (-2.94, 2.94) [0.00, 5.00] Games: 804 W: 333 L: 176 D: 295 Ptnml(0-2): 10, 61, 144, 136, 51 | 4kb threat inputs net after session 0 (base dataset 2 epoch vs 1 epoch) |
| Warreg | Pea | master | diff | 8.0+0.08 | Elo: 89.11 +- 10.60 (95%) [N=2000] Games: 2000 W: 854 L: 352 D: 794 Ptnml(0-2): 11, 121, 362, 367, 139 | progtest after training sesh 4 vs 9.1 |
| Warreg | Pea | 4kb-session4 | diff | 8.0+0.08 | LLR: 2.95 (-2.94, 2.94) [0.00, 5.00] Games: 2976 W: 826 L: 690 D: 1460 Ptnml(0-2): 35, 308, 688, 400, 57 | Test after training session 4, 80 million positions |
| Warreg | Pea | master | diff | N=5000 | Generated 2000622/2000000 Games Elo: -0.21 +- 0.45 (95%) [N=2000000] Games: 2000622 W: 869615 L: 870818 D: 260189 | Datagen for training session 4 |
| Warreg | Pea | master | diff | 8.0+0.08 | Elo: 78.84 +- 14.74 (95%) [N=1000] Games: 1004 W: 410 L: 186 D: 408 Ptnml(0-2): 5, 67, 193, 173, 64 | Test vs 9.1 after 3 training sessions, totalling around 200m positions |
| Warreg | Pea | 4kb-session3 | diff | 8.0+0.08 | LLR: 2.95 (-2.94, 2.94) [0.00, 5.00] Games: 776 W: 271 L: 131 D: 374 Ptnml(0-2): 4, 56, 152, 148, 28 | Test after training session 3, 60 million positions |
| Warreg | Pea | 4kb-session2 | diff | 4.0+0.04 | LLR: 3.00 (-2.94, 2.94) [0.00, 5.00] Games: 1482 W: 515 L: 370 D: 597 Ptnml(0-2): 19, 142, 312, 211, 57 | test after training session 2, 80m positions |
| Warreg | Pea | master | diff | N=5000 | Elo: -0.00 +- 0.00 (95%) [N=1000000] Games: 1000040 W: 342322 L: 342322 D: 315396 Ptnml(0-2): 0, 0, 500020, 0, 0 | data gen for session 2 |
| Warreg | Pea | 4kb-session1 | diff | 4.0+0.04 | LLR: 3.08 (-2.94, 2.94) [0.00, 5.00] Games: 1802 W: 660 L: 490 D: 652 Ptnml(0-2): 50, 171, 338, 243, 99 | first training session from 9.1 |
| Warreg | Pea | threat-inputs-unbucketed | diff | N=100000 | LLR: 2.97 (-2.94, 2.94) [0.00, 5.00] Games: 1546 W: 473 L: 335 D: 738 Ptnml(0-2): 23, 127, 348, 239, 36 | threat input net 4x768+4096 |
| Warreg | Pea | threat-inputs | diff | N=100000 | LLR: 3.02 (-2.94, 2.94) [0.00, 5.00] Games: 1728 W: 518 L: 382 D: 828 Ptnml(0-2): 16, 169, 382, 257, 40 | first try at threat inputs |
| Warreg | Pea | threat-inputs | diff | N=40000 | LLR: 2.99 (-2.94, 2.94) [0.00, 5.00] Games: 2758 W: 868 L: 716 D: 1174 Ptnml(0-2): 56, 284, 578, 374, 87 | 4096 threat inputs vs none with same training params |
| Warreg | Pea | threat-inputs | diff | N=40000 | LLR: 3.07 (-2.94, 2.94) [0.00, 5.00] Games: 1456 W: 498 L: 343 D: 615 Ptnml(0-2): 27, 126, 302, 211, 62 | 4096 TI per bucket |
| Warreg | Pea | horizontal-mirroring | diff | 8.0+0.08 | LLR: 2.98 (-2.94, 2.94) [0.00, 5.00] Games: 5784 W: 1742 L: 1576 D: 2466 Ptnml(0-2): 132, 646, 1202, 748, 164 | selfplay net with 0kb and mirroring |
| Warreg | Pea | horizontal-mirroring | diff | 8.0+0.08 | LLR: 2.96 (-2.94, 2.94) [0.00, 5.00] Games: 2090 W: 584 L: 455 D: 1051 Ptnml(0-2): 10, 230, 466, 299, 40 | horizontal mirroring implemented new net with 4 buckets (x2) |
| Warreg | Pea | faster-inference-from-kb | diff | 8.0+0.08 | LLR: 2.97 (-2.94, 2.94) [0.00, 5.00] Games: 1980 W: 517 L: 395 D: 1068 Ptnml(0-2): 13, 179, 492, 285, 21 | new inference that I built for kb seems to be faster |
| Warreg | Pea | kingbuckets-better-finny | diff | 8.0+0.08 | LLR: 3.00 (-2.94, 2.94) [0.00, 5.00] Games: 2342 W: 667 L: 529 D: 1146 Ptnml(0-2): 32, 228, 519, 354, 38 | make it more lazy! |
| Warreg | Pea | kingbuckets-better-finny | diff | 8.0+0.08 | LLR: 2.99 (-2.94, 2.94) [0.00, 5.00] Games: 11238 W: 2623 L: 2465 D: 6150 Ptnml(0-2): 81, 1271, 2768, 1407, 92 | make it more lazy! |
| Warreg | Pea | kingbuckets-finny | diff | 4.0+0.04 | LLR: 2.98 (-2.94, 2.94) [0.00, 5.00] Games: 1968 W: 513 L: 388 D: 1067 Ptnml(0-2): 12, 187, 479, 276, 30 | finny tables implemented? idk |
| Warreg | Pea | kingbuckets | diff | 8.0+0.08 | LLR: 2.96 (-2.94, 2.94) [0.00, 5.00] Games: 3002 W: 744 L: 617 D: 1641 Ptnml(0-2): 22, 305, 730, 412, 32 | no refresh if no bucket change |